Monitoring and Control of the Cardiovascular System During Indoor Exercise

Transcription

1 Monitoring and Control of the Cardiovascular System During Indoor Exercise A thesis written by Kaili Weng under the supervision of Dr. Steven Su in fulfilment of the requirements of the degree of Doctor of Philosophy Faculty of Engineering and Information Technology University of Technology Sydney, Australia 2012

2 CERTIFICATE OF AUTHORSHIP/ORIGINALITY I certify that the work in this thesis has not previously been submitted for a degree nor has it been submitted as part of requirements for a degree except as fully acknowledged within the text. I also certify that the thesis has been written by me. Any help that I have received in my research work and the preparation of the thesis itself has been acknowledged. In addition, I certify that all information sources and literature used are indicated in the thesis. -e. Kaili Weng

3 Abstract The increase in obesity and diabetes is of great public health, social and economic concern worldwide. Modem treadmill systems can provide effective, safe and practical indoor exercise for the consumption of extra energy. However, an uncontrolled treadmill can cause excessive exertion on the cardiovascular system. To avoid excessive cardiovascular stress, an efficient way of monitoring and controlling of exercise strength is to regulate treadmill speed and/or gradient to stimulate the exerciser's heart rate following a predefined profile. In this thesis, an automated treadmill system has been developed, which includes wireless portable ECG and tri-axial accelerometer sensors, and a Labview based control module. Based on this automated system, efficient rate detection techniques have been developed by using the pitch estimation method. Different types of multiloop integral control configurations have been proposed and implemented to regulate the heart rate and/or step rate by manipulating treadmill speed and/or gradient. These control structures have been placed under real time testing which includes Single- Inpu Single-Output (SISO), Multiple-Input Single-Output (MISO) and Multiple- Input Multiple-Output (MIMO) control by using the established Labview module. It has been found that MISO control is the most efficient method, and would be effective in making the treadmill exercise more reliable and safer in rapidly tracking the heart rate profile to achieve desired exercising outcome. For this reason, this thesis also proposes the concept of Multi-loop Integral Controllability (MIC) and proves the existence of multi-loop integral controllers which can obtain unconditional multi-loop stability of the Two-Input Single-Output automated treadmill system. The benefit of our automated control system includes assisting patients in postcardiac attack rehabilitation and therapy to safely control the heart rate to follow a suitable profile. This reduces the need for supervision by medical professionals. Furthermore, in athletics and fitness applications, an automatic control system can allow users to optimize their training intensity.

4 Acknowledgements First and foremost I would like to convey my many thanks to the remarkable people who provide support, motivation, knowledge, friendship and love on my journey to pursuit a Ph.D. Special recognition must be given to my supervisor, Dr Steven Su for his expertise, supervision and unconditional encouragement throughout the course of this project. I am particularly grateful for his wisdom that has inspired me to be a life-long learner and a passionate seeker of knowledge, and for his always being the first subject to test our new setup instrumentations, even though some of them were quite tough. I thank him with the deepest of respect. I would like to express tremendous gratitude towards my colleague Andrew Zhang for the great discussions about the research and help in many cases. I also acknowledge the support received by Sam Cheng and Sam Chui for taking time to give insights into my thesis, as well as give me some suggestions for the research. I am indebted to all of the volunteer subjects, who have given their precious time to participate in the experiment and helped me to collect the data. Special thanks to Basil Turk and Luis Dolores who were always willing to help out when I needed subjects for testing. I wish to acknowledge the important support from Australian Postgraduate Scholarship which enabled me to pursue my higher degree at University of Technology Sydney. Finally and most importantly, I would like to thank my family - my mother Chunchun Zhang for her care and being backbone during this period of my life. Thanks also go to my relatives in China. Without their eternal love and support, I would not be who I am today. 11

5 Content Abstract i Acknowledgements ii Contents iii List of Figures viii List of Tables xii Acronyms and Abbreviations xiii 1. Introduction HRC treadmill exercise Research areas in HRC treadmill exercise development Thesis objectives Main contributions of the thesis Publications Structure of tl1e thesis Background Regular physical activity Cardiovascular physiology Structure of the heart Cardiac cycle Electrocardiogram Cardiovascular exercise Cardiac rehabilitation Fitness development lll

6 Content iv 2.4. Walking pace Step rate Summary Treadmill system design Introduction Track.master TMX National instruments Labview Alive wireless monitor ECG electrodes Pedometers Bluetooth Serial protocol Data acquisition VI Communication protocol VI Data transmission VI Data storage VI HR and SR signal detection AMDF pitch detection algorithm ACF pitch detection algorithm Detection methodology Detection experimental results Summary

7 Content v 4. Labview implementation Introduction Pan and Tompkins method Band-pass filter Derivative action Squaring Moving window integrator HR sensor Peak detector block BPM mean value block SR sensor SR sensor accuracy tests Sum1nary Control configuration PID control design introduction Open loop modelling Open loop HR response HR open loop response with speed HR open loop response with gradient Open loop HR response discussion... 68

8 Content vi 5.3. Open loop SR response SR open loop response with speed SR open loop response with gradient Open loop SR response discussion Controller values HR controller values SR controller values SISO modelling methodology HR SISO control HR SISO control with speed test results HR SISO control with gradient test results SR SISO control SR SISO control with speed test results SR SISO control with gradient test results Summary MISO control methodology MISO control modelling MISO control test results SISO control and MISO control comparison... 87

9 Content vii 6.4. MISO control in exercise zones Moderate Activity exercise zone Weight control exercise zone Aerobic cardio endurance exercise zone Summary MISO analysis Introduction Multi-loop integral controllability Multi-loop integral controllability analysis for HR response Experiment design Illustrative simulation study s um.mary Mll\10 control n1ethodology MIMO control modelling Test 1 results Test 2 results s ummary Conclusions and future work Conclusions Direction for future research Bibliography

10 List of Figures 2.1: The structure of the heart and its main components (Shirley, 2007) : Chest leads: Vl, V2, V3, V4, VS and V6 (Shirley 2007) : Electrocardiogram (McGill, 2000) : Cardiovascular exercise (McGill, 2000) : Hardware and software system design : Trackmaster FVX325 series treadmill : Alive technologies wireless monitor : The orientation of the x, y and z axes in build in Tri-Axial Accelerometer : ECG Electrodes : Pedometer : USB Bluetooth dongle : USB to RS-232 connector : ECG signal : Front panel of the treadmill initialisation : Block diagram of the data acquisition l 2: Bluetooth open and close connection functions : Implementation of the open connection function : Packet layout from ECG alive technologies health monitor system : Verifying the validity of data packets : Separation of ECG and Accelerometer data : The data storage VI Block diagram of the proposed approach : Description square wave matching approach : ECG signal with different thresholds : TA signal processing by using AMDF : TA signal processing by using ACF Vlll

11 List of Figure ix 3.23: Noise polluted TA signal processing by using AMDF : Noise polluted TA signal processing by using ACF : Noise polluted TA signal processing by using AMDF+ACF : Heart rate detection : Sequence of events for the implementation of the HR detection VI : Communication to the Bluetooth protocol : Original ECG signal of a jogging test subject : Output from the band pass filter : Output from the derivative filter : Output from squaring operation : Output from moving window integrator Example of the output of the moving window integrator : Pan and Tompkins method : Pan and Tompkins heart rate detection : Peak detector block : Peak detection : BPM testing : Mean value block : Comparison open loop BPM mean response vs. Instantaneous BPM : Overview of step rate sensor : Step rate sensor initialisation : Mean calculator in step rate sensor : HR open-loop response with speed : HR open-loop response with gradient : SR open-loop response with speed : SR open-loop response with gradient : SISO PI closed loop control using speed as controlling signal... 75

12 List of Figure x 5.6: Implementation of SISO PI controller in Labview : HR SISO control with speed control response : HR SISO control with gradient control response : SR SISO control with speed control response : SR SISO control with gradient control response : Block diagram of the MISO control system : Implementation of MISO controller in Labview : Comparison of SISO speed test with MISO control : Comparison of SISO gradient test with MISO control : SISO and MISO control comparison : Exercise zones for different exercise purpose : MISO control response in the warm up zone : MISO control - speed control effort in the warm up zone : MISO control - gradient control effort in the warm up zone : MISO control response in the weight control zone : MISO control - speed control effort in the weight control zone : MISO control - gradient control effort in the weight control zone : MISO control response in the areobic cardio endurance exercise zone : MISO control - speed control effort in the areobic cardio endurance exercise zone : MISO control - gradient control effort in the areobic cardio endurance exercise zone : Multi-loop integral controllability for a 2ISO system : The 2ISO Hammerstein system : Steady state response of heart rate : The function u 1 '= (u/) : Simulation results - adjusting speed only : Simulation results - adjusting speed and gradient

13 List of Figure xi 8.1: 8.2: 8.3: 8.4: 8.5: 8.6: 8.7: 8.8: 8.9: The structure of the MIMO control MIMO control - treadmill operating system MIMO control - control algorithm MIMO control test 1 - control response MIMO control test 1 - HR controller control effort MIMO control test 1 - SR controller control effort MIMO control test 2 - control response MIMO control test 2 - HR controller control effort MIMO control test 2 - SR controller control effort

14 List of Table 2.1: Cardiac cycle of the human heart (Anonim 2000) : 3.2: The treadmill specifications The Alive wireless monitor specifications : 4.2: 4.3: Treadmill speed at 3km/h Treadmill speed at 5km/h Treadmill speed at 7km/h : 5.2: 5.3: 5.4: 5.5: Key open loop response parameters Results of HR SISO control with speed Results of HR SISO control with gradient Results of SR SISO control with speed Results of SR SISO control with gradient : 6.2: 6.3: 6.4: SISO and MISO control comparison table Performance results of MISO control in warm up zone Performance results of MISO control in weight control exercise zone Performance results of MISO control in aerobic cardio endurance zone : 7.2: Subjects characteristics Heart rate response at steady state : 8.2: 8.3: 8.4: MIMO control test 1 HR controller performance MIMO control test 1 SR controller performance MIMO control test 2 HR controller performance MIMO control test 2 SR controller performance XU

15 Acronyms and Abbreviations Units of Measure hr km/h min m sec Hour(s) Kilometres per hour Minute(s) Meter Second(s) Acronyms 2ISO ACF AMDF BPM DAQ DIC DUS ECG GAS GUI HR HRC Kp Ki Kd LES MISO Two Inputs and Single Output Autocorrelation Function Average Magnitude Difference Function Beats Per Minute Data Acquisition Decentralized Integral Controllability Decentralized Unconditional Stabi1ity Electrocardiograph Globally Asymptotically Stable Graphical User Interface Heart Rate Heart Rate Controlled Proportional Gain Integral Gain Derivative Gain Locally Exponentially Stable Multi-Input Single-Output Xlll

16 Acronyms and Abbreviations xiv MIMO MIC NI PC PID PPM RMS SISO SPM SR TA UTS VI Multi-Input Multi-Output Multi-loop Integral Controllability National Instruments Personal Computer Proportional Integral Derivative Paces per Minute Root Mean Square Single-Input Single-Output Steps per Minute Step Rate Tri-axial Accelerometer University of Technology Sydney Virtual Instrument